# RF power measurement

I'm an electronics engineer, I have primary only worked with digital and analog design, never with RF design. But finding the field interesting, I wanted to do a project mainly to get better RF design skills. So I decided to build a simple spectrum analyzer, using a PLL with integrated VCO, a mixer and some matching circuits. The signal to analyze is feed into the mixers RF port, the PLL(VCO) generates the LO signal for the mixer, and by filtering the IF output signal and measuring the power from DC to some frequency, I should be able to get the frequency content at the LO frequency. The PLL(VCO) frequency is set using a MCU, so the idea is to just set the frequencies of interest, and measure the power at the mixer IF port from DC to some frequency.

Hope that makes sense. :-)

Now how do I measure the power in the IF signal, from say DC to 10KHz?

-
If it is the same question you can just edit the previous question with more detail. – Kellenjb Jul 17 '11 at 0:21
I know and I also did that, but I think my question was so old nobody was reading it anymore. – JakobJ Jul 17 '11 at 4:27
@JacobJ - When you edit a question it tops the list of active questions again, so it should get enough attention. – stevenvh Jul 17 '11 at 6:10
Yeah. I just discovered that. :-). But anyways for some reason nobody still did not give any new answers. So just dicided to totally rewrite the question – JakobJ Jul 17 '11 at 6:15

Caveat: I have not tried this myself -- by the time you get this to work you will know more about it than I do.

Instead of putting your IF at DC, put it at some fixed frequency, say 1 MHz. Now you can run the signal through a a narrow bandpass filter to select out just the IF signal you want without worrying about DC drift issues. Connect the filtered signal to an rf detector (see also here and here). Now low-pass filter and amplify the detector output and use an ADC to measure the signal. This output will (within a certain range of input power) be proportional to the signal power you are trying to measure.

Obviously there's more to it than that; if there wasn't Agilent, Anritsu, and R&S wouldn't be able to charge 6 figures for a spectrum analyzer. But this ought to give you something to experiment with.

Edit

After consulting this book I realize that a 1 MHz IF is probably not a good idea. It is preferable to use an IF outside the input band of the analyzer, to facilitate image rejection. For example, if you were mixing a 100 MHz signal down to 1 MHz using a 101 MHz LO, you'd also get any 102 MHz signals mixed down into your IF band. In the early-70's spectrum analyzer architecture discussed in the book they actually use 3 or 4 stages of up- and down-mixing to different IF's to achieve different filtering requirements. Their first stage mixes up above the input band to enable the best image rejection.

-

Hamag made a successful range of low cost commercial spectrum analyzer using the above principle. Basically, it use mass produced TV tuner components/design and apply it in slightly different manner. See http://www.jbtech.de/measuring/hameg/hm5005.html

-